Print PageSpatial selection promotes population persistence via short-distance dispersal interacting with spatially heterogeneous risk such that offspring tend to remain near successful mothers thus increasing offspring survival rates and generating realized population growth rates significantly higher than the average, non-spatial expected growth rate. In previous simulation models spatial selection occurred with static risk patterns between generations and did not occur when risk patterns randomly re-assorted between generations. In the real world the dynamics of risk patterns likely fall somewhere between fixed and random, with patterns of risk changing gradually over many generations and having areas with both fixed and rapidly changing patterns. To investigate how spatial selection might work with more realistically dynamic patterns of risk, we developed an individual-based simulation model composed of 1) a risk module that generated spatially explicit patterns of risk with varying temporal dynamics ranging from fixed patterns through to new, random patterns each generation and 2) a predator-prey module that simulated mortality due to local predation risk and dispersal of surviving prey species. The predator-prey module was parameterized to white-footed mouse (Peromyscus leucopus) and gypsy moth (Lymantria dispar) in oak forests of the US northeast with the exception of dispersal which was varied. Parameters of the risk module (variation in risk, patchiness of risk, rate of change in spatial pattern) were varied over a wide range.
Results/Conclusions
Simulations suggest that spatial selection is not restricted to fixed risk patterns. Dynamic risk with relatively low rates of change can also generate population persistence via spatial selection. The actual rate of change were population persistence due to spatial selection was still possible depended highly upon 1) patchiness of risk with highly aggregated patterns allowing higher rates of change, 2) scale of dispersal with longer dispersal allowing higher rates of change until dispersal became so long that spatial selection no longer functioned, and 3) fatter tailed dispersal kernels allowing higher rates of change. Further, inclusion of areas with fixed patterns of risk in a broader dynamic risk landscape increased population persistence and the greater the proportion fixed risk patterns the stronger the spatial selection effect and the higher population persistence. Our findings suggest that spatial selection can play a role in real systems with dynamic patterns of risk given the rate of change in risk pattern is offset by the appropriate combinations of risk patchiness and dispersal behavior.
